A major goal of our laboratory has been and continues to be the application of recombinant DNA technology to the development of an effective live influenza virus vaccine. A suitable donor strain must meet defined criteria of reduced virulence, lack of transmissibility, and stability. Such as master strain would be used as a source of attenuating genes for newly emerged virulent antigenic variants. We attempted to rescue the NA gene DNA cloned in both the (+) and (-) sense orientation in the late region of our pSV2330 shuttle vector. This vector contains SV40 late 19S and 16S mRNA splice sequences and transcription termination sites. We succeeded, however, in only isolating antigentic variants of the N1 NA that had escaped neutralization. Data from Dr. Webster's laboratory showed that such variants within a single epitope occur with a frequency of 10 to the-5 in vitro. The "rescue" event in this experiment entails the statistically much less likely acquisition by replicating WSN virus of N2 NA virion RNA derived from cloned influenza DNA. The conservation of a 12-base sequence at the 3'-end and a different 13-base sequence at the 5'-end of all influenza A vRNAs suggest that the termini are crucial to certain aspects of viral replications, perhaps recognition by viral polymerase(s) for the production of full-length (-) strands or perhaps recognition for packaging by nucleoprotein. In the foregoing attempts at rescue, the influenza sequences were transcribed as part of a "fusion" mRNA that included SV40 sequences both upstream and downstream from the NA sequence. Thus the common termini in influenza vRNA were not available as free ends to the viral replication complex.